Method and apparatus for generating learning data for neural network

Abstract

A method for generating learning data for the neural network may comprise generating a license plate image by combining a background image, a frame image and a text image, generating a transformed image by performing at least one of a geometry transformation and a filter transformation on the license plate image, setting a text corresponding to the text image as target data for the transformed image, and generating the learning data including the transformed image and the target data.

1 . A method for generating learning data for a neural network, comprising:
 generating a license plate image by combining a background image, a frame image, and a text image;   generating a transformed image by performing at least one of geometry transformation and filter transformation on the license plate image;   setting a text corresponding to the text image as target data (label data) for the transformed image; and   generating the learning data including the transformed image and the target data.   
     
     
         2 . The method for generating learning data for a neural network according to  claim 1 , wherein the generating a license plate image comprises:
 loading one image from each of a background image group including a plurality of background images, a frame image group including a plurality of frame images, and a text image group including a plurality of text images; and   generating the license plate image by combining the loaded images, and   generating a plurality of the license plate images by repeating the steps of generating a license plate image by loading the image and combining the loaded images.   
     
     
         3 . The method for generating learning data for a neural network according to  claim 2 , wherein the loading the image comprises loading a frame image and text image, which correspond to each other in standard, from each of the frame image group including frame images having various standards and the text image group including text images having various standards. 
     
     
         4 . The method for generating learning data for a neural network according to  claim 2 , wherein the generating a plurality of the license plate images comprises
 comparing the number of generated license plate images that correspond to each text with a pre-set value; and   repeating the generation of the license plate images when the number of the license plate images that correspond to each text is less than the pre-set value, and stopping generating the license plate image when the number is greater than or equal to the pre-set value.   
     
     
         5 . The method for generating learning data for a neural network according to  claim 1 , wherein the generating a transformed image performs the geometry transformation on the license plate image such that the frame image and the text image correspond to each other in the license plate image. 
     
     
         6 . The method for generating learning data for a neural network according to  claim 1 , wherein the generating a transformed image performs the geometry transformation only on the frame image and the text image while maintaining the background image in the license plate image. 
     
     
         7 . The method for generating learning data for a neural network according to  claim 1 , wherein the geometry transformation includes at least one of length transformation, tilt transformation, and motion blur transformation. 
     
     
         8 . The method for generating learning data for a neural network according to  claim 7 , wherein the generating a transformed image performs the length transformation on the license plate image by adjusting the length in at least one direction. 
     
     
         9 . The method for generating learning data for a neural network according to  claim 7 , wherein the generating a transformed image comprises
 performing the motion blur transformation on the frame image by displaying the trace of the movement of the frame image while moving the frame image on the background image in one direction; and   performing the motion blur transformation on the text image such that it corresponds to the frame image on the transformed frame image.   
     
     
         10 . The method for generating learning data for a neural network according to  claim 1 , wherein the filter transformation includes at least one of sharpness transformation, brightness transformation, chroma transformation, contrast transformation, color transformation, noise transformation, transparency transformation, and climate application transformation 
     
     
         11 . The method for generating learning data for a neural network according to  claim 1 , wherein the method for generating learning data for the neural network further comprises training the neural network such that the target data is output through an operation on the transformed image in the neural network. 
     
     
         12 . An apparatus for generating learning data for a neural network, comprising:
 a memory in which at least one program is stored; and   a processor for executing said at least one program to generate learning data for training the neural network,   wherein the processor generates a license plate image by combining a background image, a frame image, and a text image, generates a transformed image by performing at least one of geometry transformation and filter transformation on the license plate image, sets a text corresponding to the text image as target data for the transformed image, and generates the learning data including the transformed image and the target data.   
     
     
         13 . A computer readable storage medium recording a program performing the method according to  claim 1 .